Face Search API

Upload a face, find every photo of that person

Search faces in your library/database. Works across angles, lighting, and expressions. No ML pipeline required.

Try Vecstore Free

Query

Face search query - portrait photo

"Find this person"

1 face detected

Results

Match - different angle
Side angle96.2%
Match - different lighting
Low light94.8%
Match - different expression
Smiling91.5%
Match - group photo
Different scene89.1%

1

API call

0

ML models

0

config

Use Cases

Built for apps where finding people matters.

<180ms

Search latency

99.9%

Uptime

Millions

of images searchable

Photo libraries

40,000 photos, one question: "Show me every photo of my daughter." Face search answers it instantly. No tags, no albums, no scrolling.

Works across years of photos
Multiple faces per image

Event photography

Wedding, conference, concert. Each guest uploads a selfie and gets back every shot they appear in. Replaces hours of manual sorting with one API call.

Thousands of photos per event
Zero tagging needed

Dating and social apps

Catch duplicate accounts and fake profiles. Search uploaded photos against your existing user base to flag repeated faces across accounts. Automated trust and safety.

Vectors only, no face images stored
Privacy-first architecture

Security and access

Match a captured face against an authorized database in real time. Not face recognition (putting a name to a face) but face matching (finding a match in a known set).

Real-time match and response
Configurable similarity threshold
Privacy First

No face images stored. Vectors only.

Step 1

Face detected

Vecstore finds the face in your image.

Step 2

Converted to a vector

The face becomes a set of numbers. Only the vector is stored.

Step 3

Face image discarded

The original face is never saved. You can't reconstruct it from the vector.

<180ms

Search latency

99.9%

Uptime

Any scale

Millions of images

Build vs. Buy

Skip the ML pipeline. Ship face search today.

Building it yourself
Face detection model

MTCNN, RetinaFace, or similar

Face embedding model

ArcFace, FaceNet, or similar

GPU servers

$2-10K/month in cloud compute

Vector database

Pinecone, Qdrant, Milvus, or pgvector

Sync pipeline

Queue, worker, monitoring

Estimated time to production

2-4 months

Using Vecstore
Insert images via API

Faces detected and indexed automatically

Search by face in one call

Results with similarity scores

No infrastructure to manage

No GPUs, no models, no vector DB

Privacy by design

Vectors only, no face images stored

Pay per operation

Scale up or down, no commitment

Time to production

An afternoon

Only with Vecstore

Face search + image search + text search, all from the same database.

One API key. One database. Search by face, by image similarity, by text description, or by text inside images. No separate systems.

Built-in NSFW detection across 52 safety categories.

If your platform accepts user-uploaded images, content moderation is one API call away. No separate moderation pipeline needed.

Add face search to your appin minutes, not months.

1M+ searches powered by Vecstore this year

Sign up for Vecstore
Start for Free

25 Free credits. No credit card required.